订阅最新相关文献

A classification model training method and apparatus, and a computer device, a storage medium and a program product. The method comprises: acquiring unlabeled first image data and labeled second image data; inputting the first image data and the second image data into a preset classification model, so as to obtain a first classification result of the first image data, a first classification sub-model corresponding to the second image data, and third image data, which are output by the preset classification model; on the basis of the first classification result and a second classification result that is output by the first classification sub-model, determining a first loss, and on the basis of the third image data and the first image data, determining a second loss; and on the basis of the second loss and the first loss, training the preset classification model to obtain a target classification model, wherein the third image data is an image obtained after performing mask processing and inpainting processing.

The present invention relates to the technical field of dimension identification, and in particular to a slab dimension identification method and apparatus, and a device and a medium. The method comprises the following steps: initialization settings, involving: acquiring length ratios Rn between slab surface lengths of a known slab at n known photographing distances hn and corresponding pixel lengths of n slab surface images, and acquiring width ratios R'n between slab surface widths of the known slab at the n known photographing distances hn and corresponding pixel widths of the n slab surface images, wherein n≥2; and associating an n-th known photographing distance hn with a corresponding n-th length ratio Rn and n-th width ratio R'n. The present invention combines weighing and photographic measurement techniques, and can automatically correct photographing distances between a photographed surface of a slab and a camera, so as to calculate the dimensions of the slab on the basis of the ratios of the dimensions at the corrected photographing distances to image pixels, thereby achieving higher accuracy. In addition, dimension information of the slab can be automatically measured without manual intervention, thereby improving the warehousing efficiency of the slab.

The disclosure relates to a method for prediction of skin tone perception, the method comprising: assessing L*and b*of skin of an area of interest of a user, wherein L*indicates skin lightness, b*is a colorimetric parameter indicating yellow color; calculating a perception of skin yellowness level according to the assessed L*and b*, wherein L*is negatively correlated with the perception of skin yellowness level and b*is positively correlated with the perception of skin yellowness level; and predicting the skin tone perception based on the perception of skin yellowness level. The disclosure also relates to a method for predicting efficacy of a cosmetic product, comprising: calculating a first perception score S1 of the skin of a user before using the cosmetic product and a second perception score S2 of the skin after using the cosmetic product for a predetermined period of time, utilizing the method for prediction of skin tone perception; calculating a score difference ΔS, wherein ΔS = S1-S2; and predicting efficacy of the cosmetic product based on the score difference ΔS, wherein the cosmetic product is predicted to have a perceivable efficacy if ΔS is larger than or equal to a threshold, wherein the threshold indicates a minimum perceivable change on skin tone.

An information processing apparatus includes processing circuitry configured to extract an image feature from a character image, and estimate a character string from the image feature and a number of characters.

The invention relates to a filtering method for object and edge detection in images that incorporates image processing and computer vision capabilities, enabling easier identification of edges and boundaries in images, and can be used in a variety of fields such as medical imaging, satellite and aerial imaging, security and surveillance systems, forensic science, autonomous vehicles, augmented reality, agricultural productivity analysis, land classification, art and design, military applications, industrial quality control, and product efficiency.

Systems and associated methods implement or perform receiving, in real-time, an image acquired by an imaging device, determining a symbol size for a symbol in a field-of-view of the imaging device, determining a symbol type for the symbol, comparing the symbol size to a predetermined size range, the predetermined size range being based on the determined symbol type, and providing a feedback based on the comparing, wherein the feedback indicates a relationship between an actual distance between the imaging device and the symbol and a target distance range between the imaging device and the symbol.

Provided is an input device 10 for character codes made from 27 types of character symbols, each composed of any three character elements selected from three different types of character elements and arranged in a vertical line, and each character symbol being assigned one of 26 letters of the alphabet and a space symbol. The input device 10 includes at least one processor 11 and a display 12. The processor 11 displays, on the display 12, a keyboard in which a plurality of input key columns are arranged in parallel in a horizontal direction, each input key column being composed of three input keys arranged in a vertical line for inputting one character symbol, allows a user to specify an input key, and displays, in the input key, three types of character elements in a cyclically alternating manner every time an input key is specified.

Estimating a material property parameter of fabric involves receiving information including a three-dimensional (3D) contour shape of fabric placed over a 3D geometric object, estimating a material property parameter of the fabric used for representing drape shapes of 3D clothes made by the fabric by applying the information to a trained artificial neural network, and providing the material property parameter of the fabric.

An information processing device according to one aspect of the present disclosure includes: a reception means for receiving a video; a person extraction means for extracting a person appearing in the video; and a tracking means for instructing a vison and language model (VLM) to track the person on the basis of a tracking result received from the VLM with respect to the video in a first time zone and the video in a second time zone before the first time zone. The information processing device receives, from the VLM, the tracking result with respect to the video in the first time zone based on the tracking result in the second time zone.

[Problem] To automatically read characters in document data and make it easier to provide users with information that matches their intentions. [Solution] The document service system (1) comprises a model storage unit (10) and a processing unit (20). The model storage unit (10) is configured to automatically read characters in document data (4) acquired from a user terminal (3) when it is input, and to extract and output a plurality of candidates that can be selected as the content of a predetermined item from the read characters. The processing unit (20) acquires a plurality of candidates that can become the content of a display item by inputting the document data (4) acquired from the user terminal (3) into the model, and displays the acquired plurality of candidates on the screen of the user terminal (3) in a manner that allows the user to select. [Selected Figure] Figure 1